For example x^2 is in O(x^2) but is not in o(x^2).
774 karma · joined September 22, 2020
For example x^2 is in O(x^2) but is not in o(x^2).
Whenever I try the winning font (b612, fira mono), I always go back to Liberation Mono.
It was incredible, the accent was impeccable.
If you have to chain network requests but each request depends on results of the former ones, then async/await makes it really pretty.
I have no idea how much async/await is too much.
What I do dislike is that storage loads, network requests and other things are tied to the same type, so when you look at code, you have no idea what's happening inside those async functions, but that story is for some other time.
When code is sloppily written you realize that at some point you can't use much of the existing code outside of async.
One nice example of polluting the codebase is having async local storage. Now everything that reads from the storage needs to be annotated async and now your initialization pipeline might be insanely async. Good way to avoid it is to read the whole local storage at the beginning (having a sync interface) and then everyone just reads without await.
From recent experience, I did exactly that with storage and ended up removing thousands of async annotations that were no longer necessary.
Similar "mistakes" happen for other things and then at some point you are putting loading guards everywhere because your async operations are always awaiting and letting the UI update when it should not.
Using fp-ts there is never a need for async/await and the code is still linear, (Promise is transformed to TaskEither<Error, Result>, similar to the example in the article, although you can ignore the error and just map over the result, handling the error somewhere near the end of the chain).
Problem with async/await is that the syntax is so easy that eventually the whole codebase gets polluted by it, even parts that could have been completely separated.
From the coding perspective, the problem with promises arise when there is a need to bind the intermediate results. You can either store them in an object and return that object for the next .then(fn) call, or you can nest inside a promise. I guess nesting is what annoys people the most (pyramid).
I was fortunate enough to pass everything without additional attempts and had no job at that time. Some of my friends got in after 1 year (after you fail at some step, 2nd attempt is delayed), some never managed to complete the whole process.
I started at Upwork and was constantly writing proposals and struggling with the search to find relevant jobs.
But on Toptal, even after an abrupt end of the contract I get something else in less than a week and it's zero effort from me and all effort from the Toptal team.
Even if I get bored and want a change, announcing the end of contract turns on the avalanche of offers from the Toptal team.
It's just insanely smooth and I had no idea I'd find something like that, initially.
Bitmap version of the font, I guess.
To me it just looks like a whole new batch of assumptions. Might be fictitiously valid or not.
Is it even fine to use the distribution assumptions in the later analysis?
Looks like these assumptions combined with a higher conversion rate on day 2 for control is the main reason for the surprising result (control is obviously spread out).
Whatever the reason is for using async, I believe that the mismatch happens not because Async is optimized for I/O bound programs but because it's insanely easy to use (in Rust, Javascript).
it pollutes every function with async annotation.
it allows you to quickly write the code imperatively by just annotating a function and using the async function inside it.
it results in code that is massively polluted with async when at most places you never had to use it at all (if you just restructured the code).
having async be as verbose as possible would enforce people to not use it as much and to structure their code to have async bits nicely separated from non-async bits.
Stuff before 2010 in natural language processing is ridiculous. Dynamic programming algorithms, beam search, dependency parsing (grammar) algorithms (going from O(n^3) to O(n) with cost-sensitive algorithms), a huge focus on lexical analysis, part-of-speech, graphical models (maximum entropy, conditional random fields, etc.).
Today all of these algorithms are completely irrelevant. No one needs part-of-speech anymore, or dependency (grammar) trees, or cost-sensitive reinforcement learning reductions.
I remember being so inspired by all of the work and learned a lot, but it's quite funny how Lindy works.
I'd rather have a farm raising plants than do that kind of stuff.
https://i.imgur.com/LMjg40v.png
Not sure I remember what the bug was (probably some overflow/clipping issues when dealing with 8 bit numbers) but the original image (above) was unprocessed frame from a movie (I don't remember which one).
Once the bug was fixed, there was little difference (except for the blocking effect of quantization). Was quite surprised how such a small bug creates such drastic visual difference.
But after we moved to a different country she is a completely new person.
Cloud Atlas, Sense8, and other movies Lana made after The Matrix just feel a bit cringy and off most of the time.
Even Python has it: https://docs.python.org/3/library/multiprocessing.shared_mem...